基于AI技术的新型冠状病毒肺炎CT影像特点分析

目的:通过AI技术分析新型冠状病毒肺炎(以下简称“新冠肺炎”)的CT影像特点。方法:回顾性分析武汉火神山医院经核酸检测确诊的957例新冠肺炎患者的CT影像资料。采用联影uCT76064排螺旋CT进行胸部扫描,扫描完成后将原始数据导入联影uAI新冠肺炎智能辅助分析系统进行AI计算。观察新冠肺炎患者的CT征象,包括病灶分布、影像特征等,并对左右肺的病灶体积进行比较。采用SPSS 19.0软件进行数据分析。结果:新冠肺炎患者的CT影像主要表现为双肺弥漫分布磨玻璃密度影;患者右肺病灶体积明显大于左肺,差异有统计学意义(P=0.000<0.05),右肺下叶病灶体积明显大于其余各肺叶,差异有统计学意义(P=0.000<0.05)。结论:采用AI技术分析新冠肺炎的CT影像,可以了解肺炎病灶分布特点,判断肺炎严重程度,具有重要的临床意义。

Analysis of COVID-19 CT features based on AI technology

Objective To analyze the CT imaging characteristics of COVID-19 by AI technology.Methods Totally 957 COVID-19 patients confirmed by nucleic acid detection had their CT imaging data retrospectively analyzed in Wuhan Huoshenshan Hospital.All the patients underwent chest scan with uCT76064-slice spiral CT,then the raw data obtained were input uAI COVID-19 assisted analysis system for AI computation.CT manifestations of COVID-19 were observed,including lesion distribution and imaging features.The volume of lesions in the left and right lungs was compared.SPSS 19.0 software was used for data analysis.Results COVID-19 patients had their CT images mainly characterized by ground glass opacity distributed;the volume of lesion in the right lung was significantly larger than that in left lung(P=0.000<0.05);the volume of lesion in inferior lobe of right lung was obviously larger than those of other lobes(P=0.000<0.05).Conclusion AI technology contributes to understanding the distribution of pneumonia lesions and quantifying the severity of pneumonia when used for diagnosing COVID-19,and is worthy promoting clinically.